Comment 20 for bug 198162

I couldn't remember this discussion after a couple of years, but I was amused by the comment and depressed to see that after two (more) years things have NOT turned around. I still want alternatives to Microsoft's awful software, but the major change of the last few years is that Microsoft has apparently lost their drive to be the evil empire and Apple has taken over as the leading contender (though Google is showing increasing signs of evil ambitions and capabilities).

Comment 1: I have basically given up on Ubuntu, though I am still using OLD versions on two machines. I have experimented with later versions on several machines, and Ubuntu has become less and less ready for prime time. I managed to avoid ever buying a Vista machine, Apple has dropped out of my consideration, but I increasingly doubt any acceptable alternatives will appear before I am obliged to buy a Windows 7 box... Right now Google's Chrome looks to be the best alternative, but I feel like I've wasted a LOT of time with Ubuntu, and that casts a shadow on all of the Linux alternatives. (My company uses RHEL, which I've never managed to like. If it was actually offered as an option on a machine, I'd probably take it over Windows--but so far it has not been competitive in the real-world market.)

Comment 2: I think the REAL problem is Ubuntu's financial model doesn't provide sufficient push for regression testing against the pushes for the development of new features. The most serious problems I've encountered over the years are almost always breakage in old features. The problems with new features are annoying, but rarely show stoppers. Ergo, I suggest Ubuntu might be salvaged with an alternative funding model that supported MUCH more testing, especially boring old regression testing. Perhaps something like this:

http://eco-epistemology.blogspot.com/2009/11/economics-of-small-donors-reverse.html

In applying the model to Ubuntu's situation, I think that the project models should have substantial allocations for testing in their budgets--but the virtual shareholders would also become the highest-priority candidates to become testers. In other words, if you bought a share in a particular project, you could also volunteer as a tester at that time, and record your configuration. At that point it would be a bit of a lottery, but essentially the Ubuntu people would be picking testers to maximize coverage, and the winners would be paid for their testing work. The value paid for testing would probably be more than the cost of shares, so the winners would be happy and it would be yet another motivation to support a development project. I think the main advantage is that the projects would be buying testing from normal users--but perhaps that's because I think that the earliest users of most programs tend to be on the strange side or stranger. (The main risk might actually be a kind of gambling by people who think they have especially useful configurations for testing...)

I still want more options, and I wouldn't mind a bit if the Ubuntu people could turn things around--but I'm not worrying about it at this point, and I even resent the loss of capabilities and increasing bugginess of newer versions...